© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 13846
    Prev
    Next
    INDEX
    Explanations

    word pairs indicating negative impacts or consequences

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    277
    +0.10
    0.4%
    407
    +0.10
    0.4%
    726
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    3
    +0.10
    0.04
    669
    +0.10
    0.04
    736
    +0.09
    0.04
    Negative Logits
    involved
    -0.74
    ften
    -0.66
    hal
    -0.64
    got
    -0.62
    trap
    -0.62
    âĸĪâĸĪâĸĪâĸĪ
    -0.61
    az
    -0.61
    augh
    -0.60
    Lie
    -0.59
    roo
    -0.59
    POSITIVE LOGITS
     morale
    1.17
     livelihood
    1.05
     havoc
    1.04
     mobility
    0.89
     productivity
    0.86
     ecosystems
    0.85
     economies
    0.83
     cohesion
    0.82
     friendships
    0.81
     budgets
    0.79
    Activations Density 0.376%

    No Known Activations